product-launch-tracking
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from multiple external and untrusted sources to build context and perform analysis, creating a surface for potential injection.
- Ingestion points: The agent is instructed to fetch and research company websites and public sources, as well as process user-provided CSV exports and pasted data.
- Boundary markers: The instructions include a safety protocol in the house rules (Rule 7) requiring the agent to treat fetched content as data rather than instructions, which serves as a conceptual boundary.
- Capability inventory: The skill has the ability to fetch web content, write to local files (such as
.agents/product-context.mdand a ledger), and invoke platform tools for visual rendering. - Sanitization: No specific technical sanitization, such as schema validation or filtering for executable patterns, is mandated for the content retrieved from third-party sites.
- [DYNAMIC_EXECUTION]: The skill leverages platform-specific visualization tools to generate dynamic UI components based on the analyzed data.
- Evidence: Under the "Visual launch dial" section, the skill instructs the agent to use tools like Claude Artifacts to render progress bars and lists.
- Context: While this is a standard platform capability, the generation of visual artifacts from dynamically analyzed data constitutes a low-risk form of script/UI generation.
Audit Metadata